Penetration Testing Cost in 2026
What drives penetration testing cost, how to turn scope into tester-days, and a quote-comparison worksheet so you can tell a cheap test from a good one.
Penetration testing cost is set by one thing above all others: how many days a qualified tester spends on your systems. Everything else — the type of test, the size of the estate, the methodology, the certifications behind the tester — matters because it changes that number. If you can estimate the days, you can estimate the price, check a quote for sanity, and tell the difference between a test that is genuinely cheaper and one that is simply shorter.
This guide shows you how to build that estimate, what legitimately moves the price up or down, and how to compare quotes that look nothing alike. More background on testing sits in the ResilientX Learning Center.
Why quoted prices vary so much
Published price ranges are close to useless on their own, for three reasons.
First, almost every published figure is in US dollars and reflects the US market. European buyers procuring in euro, under European regulation, with European day rates, are reading someone else's market.
Second, the unit is undefined. A "web application test" can mean a two-day authenticated scan-and-verify or a ten-day manual test of every role and workflow. Both are sold under the same label.
Third, scope is elastic. Providers quote what you describe. Describe your estate loosely and you will get a low number that grows once testing starts.
That is why this guide does not lead with a price. It leads with effort.
How do you estimate penetration testing cost?
Use a three-part model. It is the same arithmetic any reputable provider runs internally.
Cost = (scoping days + testing days + reporting days + retest days) × day rate
Ask each provider for their day rate and their day estimate separately. A provider who will not break a fixed price into days is asking you to compare numbers that cannot be compared.
Step 1 — count your scope units
A scope unit is one thing that needs testing effort on its own:
- One web application, counted per role, not per site — a three-role app is three times the authorisation testing
- One API, counted per collection of endpoints sharing an auth model
- One external IP range, counted per /24 or per live-host cluster
- One internal network segment or VLAN
- One mobile app, counted per platform
- One cloud account or subscription
Step 2 — convert units into tester-days
Effort follows the four phases in NIST SP 800-115, Technical Guide to Information Security Testing and Assessment: planning, discovery, attack and reporting, with reporting running alongside the other three rather than after them. That structure is why a "quick test" still carries fixed overhead — planning and reporting do not shrink to zero.
| Scope unit | Typical effort | What drives it up |
|---|---|---|
| External network, one /24 | 2–3 tester-days | Many live hosts, legacy services |
| Web application, per role | 3–5 tester-days | Complex workflows, file upload, payments |
| API, per auth model | 2–4 tester-days | Many endpoints, weak documentation |
| Internal network, per segment | 4–6 tester-days | Flat networks, large AD estates |
| Mobile app, per platform | 3–5 tester-days | Hardening, certificate pinning, offline logic |
| Cloud account review | 2–4 tester-days | Multi-account estates, IaC sprawl |
Add 1 day for scoping and rules of engagement, and 1–2 days for reporting and the debrief, per engagement rather than per unit.
Step 3 — worked example
A mid-sized company wants: one external range, one customer web app with three roles, and the API behind it.
| Item | Days |
|---|---|
| Scoping and rules of engagement | 1 |
| External /24 | 2 |
| Web app, 3 roles at 3.5 days | 10.5 |
| API, one auth model | 3 |
| Reporting and debrief | 2 |
| Subtotal | 18.5 |
| Retest of fixed findings | 2 |
| Total | 20.5 tester-days |
Multiply by the day rate each provider quotes you. The point of the model is not that 20.5 is the right answer for your estate — it is that when one quote says 8 days and another says 20, you now know exactly which assumptions to challenge.
What actually changes the price?
Test type and knowledge level. A black box test spends days on discovery that a white box test skips, but white box testing covers more ground per day because the tester is not guessing. Neither is automatically cheaper — they buy different things, as black box and white box testing sets out.
Methodology. A test run to a published methodology such as the PTES methodology or OSSTMM has a defined floor of work. A test with no stated methodology has no floor, which is where unusually low quotes come from.
Manual depth versus automation. Automated scanning is cheap and finds known issues. Business-logic flaws, authorisation bypasses and chained exploits need a person. CREST's guidance on running a testing programme is blunt about the trade-off: obvious vulnerabilities surface first and each additional day yields more obscure findings, so buyers should balance risk against cost rather than simply buy the shortest test.
Objective-based red teaming. Testing against a goal — reach the payment system, exfiltrate a dataset — costs more than a scoped assessment because the tester works across the whole estate and evades detection. MITRE ATT&CK red teaming explains what that engagement actually involves.
Tester qualifications and independence. DORA Article 24(4) requires that tests be "undertaken by independent parties, whether internal or external", and that conflicts of interest are avoided where testers are internal. Independence and certification are cost drivers you should want to pay for.
Retesting. Ask whether retesting is included and for how long after the report. A report you cannot act on and re-verify is a sunk cost.
How does regulation change what you must budget?
This is the part most cost guides leave out entirely, and for European buyers it sets the floor.
Under DORA, Article 25(1) requires the digital operational resilience testing programme to provide for a range of tests explicitly including "penetration testing". Article 24(6) requires financial entities other than microenterprises to ensure that "at least yearly" appropriate tests are conducted on all ICT systems and applications supporting critical or important functions. So annual testing of critical systems is not a best practice for these entities — it is the baseline.
Article 26(1) goes further for entities identified for advanced testing: they must "carry out at least every 3 years advanced testing by means of TLPT", threat-led penetration testing, performed on live production systems covering critical or important functions. A competent authority may raise or lower that frequency. Where an ICT third-party provider is in scope, Article 26(3) requires the entity to secure their participation while retaining full responsibility.
Under NIS2, Article 21(2)(e) requires measures covering "security in network and information systems acquisition, development and maintenance, including vulnerability handling and disclosure", and Article 21(2)(f) requires "policies and procedures to assess the effectiveness of cybersecurity risk-management measures". Testing is how most entities evidence that assessment.
Practical consequence for budgeting: if you are in scope, plan a recurring annual line, and if TLPT applies, a much larger three-yearly line. Our penetration testing services cover threat-led testing built for the DORA requirement alongside standard scoped assessments — book an expert-led penetration test when you know which of the two you need.
How do you compare quotes fairly?
Put every quote through the same grid before you look at the totals.
| Question | Why it matters |
|---|---|
| How many tester-days, by phase? | The only comparable unit |
| What is the day rate? | Separates rate from effort |
| Which methodology? | Sets the floor of work |
| Manual effort versus automated share? | Tells you what you are buying |
| Tester certifications and independence? | A DORA Article 24(4) requirement |
| Is retesting included, and for how long? | Often the hidden cost |
| Is a debrief included? | Turns a report into remediation |
| Who owns the report and raw data? | Matters for audit and insurers |
CREST's procurement guidance recommends that supplier requirements be formally defined, recorded in a specification and grounded in a cost/benefit analysis, and it warns that buyers treating testing as a box to tick rarely get a useful result. If one quote is dramatically below the others, the grid usually shows why within two rows.
Where buyers waste money
- Testing everything at the same depth. Tier your estate and spend the days where compromise actually hurts.
- Scoping late. A test booked after the release date gets whatever time is left, which CREST identifies as a recurring cause of insufficient testing budget for systems under development.
- Paying for discovery you already have. Give the tester your asset inventory and architecture diagrams. Discovery days spent rebuilding what you already know are days not spent attacking.
- No retest. Findings that are never verified as fixed leave you with a report, not an improvement.
- Annual-only testing. CREST notes that any test captures only a moment in time, and recommends testing regularly rather than as a one-off exercise.
Frequently asked questions
How much does a penetration test cost?
There is no single honest figure, because the unit differs between providers. Estimate it as tester-days multiplied by the provider's day rate: a small scoped external and web application engagement typically lands in the mid-teens of tester-days once scoping, reporting and a retest are included. Ask every provider for days and rate separately, then compare the days.
Is a cheaper penetration test ever the right choice?
Yes, when the scope genuinely is smaller or the risk is lower — a static marketing site does not need the depth a payments platform does. It is the wrong choice when the saving comes from fewer manual days on a system that matters. CREST's advice to balance risk against cost is the test to apply.
Does penetration testing cost more than vulnerability scanning?
Considerably, and they answer different questions. Scanning checks whether a vulnerability might exist; a penetration test exploits it to confirm that it does and shows what an attacker reaches next. Most programmes run continuous scanning for coverage and periodic penetration testing for depth and assurance.
How often should we test?
Under DORA, financial entities other than microenterprises test ICT systems supporting critical or important functions at least yearly, with TLPT at least every three years for identified entities. Outside those regimes, test annually and after any material change to the system or its environment.
What should the price include?
Scoping and rules of engagement, the testing itself, a written report with risk-rated findings and remediation guidance, a debrief, and at least one retest of fixed findings. Confirm each of these in writing, since the last two are the ones most often quietly dropped from a low quote.
Estimate the days first. Once scope is expressed in tester-days, a quote stops being a number you have to trust and becomes one you can check.
Book an expert-led penetration test — talk to our testing team about scope and we will come back with the days, not just a price.